University of Cambridge > > Inference Group > Variational Gibbs Sampling

Variational Gibbs Sampling

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Phil Cowans.

I introduce a MCMC method for sampling from latent variable models. The sampling scheme circumvents the traditional latent variable sample by creating a transition kernel with the required parameter posterior as its invariant distribution, hoping to smooth over local maxima and trapping states in the latent variable space. In general this kernel is not analytically tractable and I approximate it with a simpler distribution using an EM bound; I’ll also discuss methods to correct this approximate chain. Finally I’ll relate the method to two-stage Gibbs sampling, EM and variational methods.

This talk is part of the Inference Group series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


© 2006-2023, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity